AI Speech Analytics System

🚀 Overview

AI Speech Analytics System is a practical demo project tailored for a Swiss company Spitch AI. It processes call audio data to provide actionable insights through:

• Accurate Transcription: Converts uploaded audio files into text.
• Extractive Summarization: Uses a controlled TextRank algorithm to extract key sentences directly from the transcription.
• Sentiment Analysis & Entity Extraction: Analyzes sentiment and extracts keywords and named entities to offer a quick overview of call dynamics.

This project demonstrates end-to-end capabilities—from speech-to-text conversion to insightful data extraction—using state-of-the-art NLP techniques.

---

📺 Demo

---

🔥 Key Features

• Accurate Transcription: Converts audio files (WAV, MP3, M4A) into text.
• Extractive Summarization: Uses TextRank to produce concise, faithful summaries without hallucinations.
• Sentiment & Entity Analysis: Provides overall tone, polarity, subjectivity, and identifies key entities.
• Punctuation Restoration: Enhances raw transcriptions by restoring punctuation for better readability.
• Interactive Dashboard: Built with Streamlit for an intuitive, real-time analysis interface.

---

🛠️ Tech Stack

---

📌 How It Works

1. Audio Processing:

   • Upload an audio file via the Streamlit dashboard.
   • The file is transcribed into text using a speech-to-text engine.

2. Data Enrichment:

   • Punctuation Restoration: The raw transcription is enhanced to improve readability.
   • NLP Analysis: Extracts keywords, performs sentiment analysis, and identifies named entities.

3. Extractive Summarization:

   • Applies TextRank (via Sumy) to select key sentences from the restored transcription.
   • Generates a concise summary that faithfully reflects the call's content.

4. Interactive Insights:

   • The dashboard displays transcription, key insights (keywords, sentiment, entities), and the extractive summary in real time.

---

🔧 Setup & Installation

# Clone the repository
git clone https://github.com/YourUsername/ai-speech-analytics.git
cd ai-speech-analytics

# Install dependencies
pip install -r requirements.txt

# Then download the Spacy English model
python -m spacy download en_core_web_sm

# Run the Streamlit UI
streamlit run streamlit_app.py